Fiza Iqbal
Full-stack dev building AI agents, voice bots & web apps
New to Contra
Fiza is ready for their next project!
StockPilot is an AI-powered inventory management desktop app I designed and built end-to-end, using NestJS and Electron. It tracks stock across the business and layers AI-driven decision support on top: low-stock alerts, restock recommendations based on demand patterns, guidance on where to invest and what to prioritize, and a churn radar that flags at-risk trends before they become a problem. It also automatically generates weekly and monthly reports, so business owners get clear insights without manually pulling numbers together. The goal was to turn inventory management from a reactive, manual process into something the AI actively helps run.
0
14
Voxa is an AI receptionist SaaS I designed and built end-to-end, on my own, to solve a problem every service business faces: missed calls mean missed revenue. It answers phone calls and WhatsApp messages 24/7, automatically books, reschedules, or cancels appointments, captures leads, and handles customer questions all through natural conversation, with no scripted menus. The system is multi-tenant, so each business can configure its own hours, services, and pricing. At its core is a tool-calling AI agent shared across both voice and WhatsApp channels, built with safeguards so the AI can never double-book a customer even if a request is retried, and a human-escalation path for anything it can't confidently resolve on its own. I built the real-time voice pipeline myself speech-to-text, AI-driven conversation handling, and text-to-speech and evaluated several telephony infrastructure options to keep the system cost-efficient and self-hostable rather than dependent on expensive third-party voice AI platforms.
0
29
Loxala is a freelance marketplace connecting clients with verified freelancers through secure escrow payments, an AI-driven "Skill Matcher" for job recommendations, and a reputation/leveling system. I worked on this as a Full-Stack Developer on the engineering team at Airbridge Devs, contributing to vendor/freelancer management features, customer flows, and admin dashboards using NestJS, React, Next.js, and PostgreSQL
0
21
VirtualTry.Lab is an AI-powered virtual try-on and styling platform I built for a fashion e-commerce brand. Users paste a product link and upload a photo of themselves, and the system generates a realistic try-on preview along with an HD render, letting shoppers see how a garment actually looks on them before buying. Behind the scenes, I built an AI-driven segmentation step that isolates clothing from a photo, a personalization layer that detects skin undertone and learns each user's taste over time, and a trend-tracking system that scores fashion trends from multiple sources. I also built an AI shopping assistant that reasons over the live product catalog to recommend outfits and answer styling questions, with a reliable fallback system so the assistant never breaks even if a component fails. The project spans roughly 13,000 lines of code across a Next.js frontend, a NestJS backend, and a Chrome extension, and required solving real-world engineering problems like handling anti-bot protections, coordinating multiple AI providers for image generation, and designing a storage layer that can scale from local disk to cloud storage.
0
34